Variable and its Types in Statistical Analysis
Published byat September 22nd, 2021 , Revised On July 5, 2022
You might have come across the word variable in algebra a million times. The case is a little different in statistics. And, you will find that out in a moment.
What is a Variable?
A variable is an attribute to which different values can be assigned. The value can be a characteristic, a number, or a quantity that can be counted. It is sometimes called a data item. To precisely say, it is anything accepting various values. Examples of variables can be gender, expenses, hair colour, number of schools in a city, and so on.
Though there are many types of variables in statistics, they are broadly divided into four categories or groups in statistics. These are:
- Quantitative Variables
- Categorical Variables
- Dependent Variables
- Independent Variables
Quantitative Variables are those variables that can be counted in terms of figures and numbers. They are also called Numeric Variables. A person’s height is one example of a quantitative variable because it can take on different values. You can be 3 ft., 5 ft., and so on.
These properties or characteristics can vary from person to person. Thus, quantitative variables possess a natural ranking or order. They are further divided into two more categories, namely Discrete and Continuous Variables.
All the values that are countable and can have a finite number of possibilities are called discreet variables. They are usually integers. For instance, citizens of a city, teachers in a school, and the number of children in the family.
Now you must be wondering how someone can count the total number of citizens in a city. Well, that is something tedious, challenging, and time-consuming, but not impossible.
Continuous variables, on the other hand, are numbers that can take forever to count. An example of a continuous variable is a person’s age or time, both of which you cannot count. Why not? Because they cannot stay fixed. A person’s age could be 35 years, two months, 12 hours, 4 minutes, 10 seconds, nanoseconds, picoseconds, and so on. However, there is a way to make it a little convenient. And that is, by first turning it into a discrete variable and then doing the measuring and counting.
Not sure if you get this?
Here’s an example that can help you comprehend.
You cannot count the age of a person or time in general, but if you change it to:
- A person’s age in 5 years
- Hours, minutes, and seconds till 10 pm tomorrow
You can get accurate values this way.
Categorical variables belong to a broad group of variables that are not numbers and can never be counted. They are also popularly known as Attribute or Quantitative Variables. They are further categorized into three types:
- Nominal Variables
- Ordinal Variables
- Dichotomous Variables
Let’s dig deeper into each of these.
Nominal variables are those variable that is used to label and group various attributes being measured. It takes qualitative values to reflect different categories, and there is no intrinsic ranking of these categories. For instance:
- Blood Type
- Zip Codes
Ordinal variables, on the contrary, are those where the order of values matters but not the difference between these values. For instance:
- Educational Level
- Hierarchy in Office Positions
- Income Level
An important thing to consider here is that difference between adjacent values does not have to have the same meaning. For example, the difference between the two salary levels, “Less than 30K” and “30K-60K,” does not have the same meaning as the difference between the two salary packages “30K-60K” and “more than 60K”.
The name is self-explanatory! A dichotomous variable has only two (di) levels or categories. For instance, if you ask somebody if they own a car or not, it would have two possible answers. Similarly, say you hear about birth in your family, then you know it would either mean a girl is born or a boy.
The independent variables, also known as resultant variables and experimental variables, are computed in research to depict the effects of dependent variables. That is why in statistics, we say that they are determinants of the value of the dependent variable.
In other words, the variations of these variables do not depend on another variable. They have measured inputs whose variation solely depends on the person dealing with the variables.
For instance, say you are required to move an object from point A to be in a given time. The independent variable, in this case, would be the speed which depends on the person running/walking, while the dependent variable is the time that changes with the change in speed of the person.
You might have guessed what these variables are by now with everything we have discussed earlier.
Well, you are right!
Dependent variables are the complete opposite of dependent variables, which means their values change with another variable. So, if the value of the independent variable changes, the dependent variable will also vary.
But how is this change determined?
The course of this change can be found out by a function representing the relationship between the independent and dependent variables. In mathematics, it is expressed as a function of the independent variable. For instance, if z =h(x) = 4x+5, then ‘z’ here would be the dependent variable and ‘x’ is the independent one.
Other types of variables you might also come across:
The latent variable is one that cannot be directly observed. They are hidden!
So, how do you detect these variables?
You can detect this variable from the effects they have on other variables. For instance, a person’s overall health is a latent variable because it can only be detected by measuring the physical properties of the body. These include the body temperature, sugar level, blood pressure, and so on.
It is the addition of several other variables in an experiment. For instance, BMI can be a composite variable as you need a person’s height and weight for it.
A confounding variable is the crux of all your problems.
Yes, you heard it right!
These are extra variables that you did not account for and can ruin the whole experiment. They trick you by suggesting a correlation between different variables when in reality, that is never the case. For instance, if you are doing some research on whether lack of exercise can lead to weight gain or not. Lack of exercise here is the independent variable, weight gain is the dependent variable, while any other variable such as the effect of weather conditions would be a confounding variable.
A discreet variable is one that you can measure of count in a finite amount of time. For instance, the number of students in your class.
Quantitative variables represent some measurements and take numerical values. For instance, the age of a person. Categorical variables, on the other hand, are not countable and place individuals into several other categories. For instance, the gender of a person. It cannot be counted; instead, it is divided into two or more labels, such as male, female and other.
A dependent variable is one directly relying on the independent variable. For instance, if you are reading a novel and have to finish it in a given timeline. Time here would be the dependent variable, and the speed or number of pages you read will be the independent variable.
A dichotomous variable is one where there are two categories, both of equal significance. For example, if your survey question says, “are you a smoker?” it could either be ‘yes’ or ‘no’.
For queries and questions, please leave a comment below and we will get back to you soon.